Andy makes some excellent points.
Comments below...
Mark
Anselm Hook wrote:
> Speaking for myself only, I'm afraid that debates between VRML and Java bore me to
>tears. Just evaluate all of the technologies
> directly and make your own decision. One is largely declarative and the other
>largely procedural. These are all just grammers -
> each trying to define the problem space of 3d graphics. They overlap in some
>degrees, like say a chassis and an engine for a car,
> but they have degrees of freedom relative to each other (at least it seems to me)
>which make it difficult to strictly rate them.
>
> If I really had to be critical I would say that it is erroneous to be fixated on 3D
>graphics at all - it's just the shallowest part
> of building a simulation.... There is an *incredible* hubris in calling VRML
>"Virtual Reality" - as if virtual reality has anything
> more than the most passing relationship to visualization. The real focus of VR is
>the dynamical simulation of an underlying
> reality.... and any transport protocol or grammer can be used to shovel the polygons
>in your direction. People doing haptics
> research probably care a lot more about millisecond response time to collision
>events than they do to the color of a polygon. Now
> if VRML had specified certain guarantees about collision and suchlike then that
>might be a different story.
That's why a general 'Component composition syntax' is crucial,and not an 'inheritance
hierarchy' or 'node/field taxonomy'- the
precise needs and scope of the 'Virtual Reality' domain of
discourse will never be fully known and finalized.
Thus, interoperability of Components is key. An attempt to create
an application domain 'Type hierarchy' or Node/field set are very
brittle and transitory - we can't anticipate our needs - that's why the
key is interoperable Components. However, in order to make it easy
to use, a declarative syntax for composing Components is also key.
The ability to use a declarative composition and 'wiring' of
predictable Components of any sort, plus the ability to encapsulate
arbitrary programming in easy-to-use focused Components, is
a very powerful combination. It is also 'future-safe' since it accomodates
the use of Components from any domain and any time of creation
I use both Java3D and VRML. Each has it's relative strengths
and weaknesses.
I think that the goal of X3D is to use what has been learned to date
about declarative syntax (VRML,'Bean Markup Language') along
with software theory -Components and Frameworks of Components
(JavaBeans, Corba Components) and create a declarative
syntax for composing interoperable Components for the
expressive purposes of interactive 3D. The functionalites used
in the composition are a matter of availability, design and aesthetics.
(It is a generally elegant and effective approach to any
application domain)
>
>
> Also, I should qualify that while I feel they are complementary I am personally
>critical of VRML. We (as a community) should really
> be defining component level architectures which are application neutral (such as
>Mark Rudolph is doing, and such as the X3D
> consortium is doing). VRML's extern proto support technically does allow third
>party components but perhaps just for reasons of
> having to shuttle back and forth between various grammers it has always been
>challenging to use. I was a big fan of Justin Couches
> VRML bindings to Java but even then it was difficult to debug all the way through a
>system and this was an impedance for me as a
> professional games developer used to having total control over a system. And recall
>the fiasco with Newfire who tried to build a
> business around VRML... it's clear that VRML (regardless of comparison to Java3D)
>has its own issues.
>
> (In my opinion) In the areas where VRML and Java3D do overlap - behaviors and
>components, Java and Java3D are more mature. Java
> itself provides a nice concept of packaging and a fairly good approach to component
>level assembly using JavaBeans... if you're a
> programmer this all forms a very nice declarative and procedural environment that is
>fairly application neutral. (But of course
> that is what a general purpose programming language is - an application neutral
>framework for hanging chunks of code from various
> vendors).
>
with the 'hanging' being made effctive and reliable to the degreethat the Component
'chunks' adhere to some set of 'meta-agreements'
such as the JavaBean 'design pattern' naming/semantic conventions,
with room for additional 'meta-agreements' such as 'method
contracts' (pre-c,post-c,inv), and eventually maybe even
usage semantics, and maybe even some sort of introspectable
'complexity' information - 'time/space' info of code etc
It seems that even the nature of our Component 'meta-agreements'
must be able to evolve to accomodate unanticipated
needs and conceptual advances. Again, a declarative syntax
such as XML provides a way forward. The two work very
well together. This is why I think they are key to X3D,
and programming and application building in general.
> Well, I'll be announcing my own Java3D project soon, and in fact I use VRML heavily
>so I shouldn't complain.
>
> - Anselm Hook
>
> -----Original Message-----
> From: Thomas Bendig <[EMAIL PROTECTED]>
> To: Anselm Hook <[EMAIL PROTECTED]>
> Cc: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
> Date: Friday, September 10, 1999 7:17 AM
> Subject: Re: [JAVA3D] Java3D vs VRML and X3D
>
> >Hi Anselm,
> >I don't think so. These technologies are targetting the same or at least very
>overlapping application areas.
> >I think, every application done in VRML could also have been done based on Java3D.
>On the other hand I think that it's no problem
> to build very complex applications in Java3D but in VRML such complex applications
>are very hard to handle because of the interfaces
> to other technologies (java, javascript, eai, ...).
> >
> >So I would say VRML is a nice geometry file format, but for building complex
>applications Java3D ist the better technology.
> >
> >And even those people who don't what to touch code (designers, architects, ...)
>could use standard Java3D applets which understand
> their specific file formats (max, 3ds, dxf, ...)
> >
> >Or am I wrong?
> >
> >Thanks for your answer.
> >
> >Thomas Bendig
> >[EMAIL PROTECTED]
> >http://www.echtzeit.de
> >
> >
> >> -----Original Message-----
> >> From: Anselm Hook [mailto:[EMAIL PROTECTED]]
> >> Sent: Donnerstag, 9. September 1999 18:16
> >> To: [EMAIL PROTECTED]
> >> Subject: Re: Re: [JAVA3D] Java3D distribution with browsers
> >>
> >>
> >> X3D and VRML are totally complementary to Java3D. They will only
> >> help each other.
> >>
> >> >
> >> >May I ask you (all) another question?
> >> >What do you think about X3D and VRML in relation to Java3D?
> >> >Will Java3D make it obsolete?
> >> >
>
> ===========================================================================
> To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
> of the message "signoff JAVA3D-INTEREST". For general help, send email to
> [EMAIL PROTECTED] and include in the body of the message "help".
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
